Seven Layers of Tactical Decision-Making

Pennsic was good, but I don't want to talk about that.

I went to practice yesterday, and it was frustrating. I couldn't figure out why things weren't working right, until I was leaving practice talking to Rowan, and I realized that I had completely forgotten to implement a coherent strategy all practice. Like, at all.

Some background - in the way I think of fencing, there are several levels to the game. They are interconnected, and the boundaries between them can be fuzzy sometimes, but they go something like this, from bottom to top:

• A physical movement.
• This is something like "extend your arm" or "step left while turning your shoulder behind you", or even could be "perform a lunge" or "execute a giarata". At this level of consideration, your concern should be primarily internal. Are you doing the thing correctly? Could you do it more efficiently? Is it happening in the way you envisioned it?
• This is where most solo-drills live.
• A technique.
• This is the point at which we consider that we have an opponent. A technique is different from a movement in that the way you perform it changes based on the positioning of yourself and your opponent. Here, we're not considering edge-cases or weirdness. We're considering performing the technique, and it going correctly.
• This is where most pair-drills aspire to be.
• A specific implementation of a technique. (I will call this an operation, henceforth)
• This is where messy stuff comes in. There are idealized versions of techniques, but there are a lot of squishy places where the technique "goes wrong", or your opponent does something unexpected. This can be anything from "oh shit, my opponent disengaged at the exact same time I did my thing" to "my opponent isn't letting me get them firmly within the bounds of the technique, so I need to figure out how to modify or adapt the technique to make it work given what they are letting me have".
• This is what we consider when our opponent is staying too far away, or not giving you as strong opposition as you want, or they are pushing on your sword way harder than you expect them to. Doing this part well in a bout tests the boundaries of your knowledge of techniques. Will this work here, or will it fail here?
• This is where I personally get stuck on period manuals. "Oh, but what if they have a weapon that's longer or shorter? How does that change things? What if they do this obscure thing? And that other one?" It's somewhat exhausting, really, and I really wish more period masters covered possible variations more thoroughly.
• This is also covered in pair-drilling, but it's hard to actually get people to focus on it when they're concerned with "doing the drill right". Frequently people think that the problem is with them, rather than trying to vivisect the technique to figure out what makes it tick.
• This is closer to what I consider an "exercise" rather than a "drill".
• The possible results of an operation. I guess you could call this "an exchange"?
• This is where one considers the places your opponent could be after you do your technique. "After you do your technique" is a vast over-simplification, though. Humans have a constant loop of perception and action going in their heads. As well, different types of perception happen faster than other types - you can react to a sound faster than to a sight, and faster still to a touch. This is neurochemical truth, and unavoidable.
• Back to the point here - there is a small gap between what you perceive and the actual state of the world. There is another small gap between the decision to perform an action, and your muscles implementing that action. This layer of the tactical process is all about considering what you can perceive during your operation, and what that could mean in tactical terms.
• I call the aggregate time of those two gaps "a tempo". Many people disagree with me, including period masters. I call it such because it is the smallest amount of time that you can be sure that your opponent will not react to your action. They might predict what you are going to do, and they might even predict when you're going to do it, but they won't REACT to the action because it is physically impossible.
• An example is in order here. I am so, so sorry.
• If you are attempting to find your opponent's blade by making contact with it on the high inside line and your blade doesn't touch it when you expect it to, what could they be doing?
• They could have yielded around, moving their hilt away from the line but keeping their tip on-line.
• They could have performed a disengage or a disengage-attack.
• They could be performing a half-disengage or low-line attack.
• Or, they could have pulled their blade back entirely, either by pulling far backwards or by performing a moulinet.
• (Or they could perform some bastard combination of the above.)
• The job we have here is to figure out, given the small amount of information we have (DID NOT FEEL SWORDS TOUCH) what we should do in order to infallibly not get stabbed. The swords-not-touching is the very first information we receive that Something Is Not Going According To Our Initial Operation. In some cases, the correct thing to do might be to wait until we see what they are doing with our eyes, instead of acting prematurely.
• I'm disregarding the idea that someone might disengage earlier because I'm currently assuming that we're acting
• In this particular case, I feel relatively comfortable (with my tip-heavy blade) doing a mid-blade rotation from my wrist to place my tip low and my hilt to the left, creating a descending cut. That rotation lets me avoid acting directly counter to the original gaining motion, which would be slow. That descending cut will catch everything except for the fourth option there, and the fourth option will take enough time to complete that I can perceive that it is happening soon enough to counter it.
• This leads to the next blossoming perception loop, in which we perceive if our opponent has been caught by our cut. And, if not, why not and what can we do about it?
• This is where tactics get interesting. As you can see from the above example, the tree of possibilities blossoms too quickly to map out exhaustively. Especially for beginners, this is where having a coherent style Matters A Lot. Most period masters are relatively congruent and cover most situations pretty fully. Even if they don't cover a specific situation, there's probably something in their manual that is relevant and can be adapted to fill the gaps. They usually aren't exhaustive in enumerating possible results, however.
• Thibault's manual covers this really, really well. In excruciating detail, really, which is why it's So Damn Long. He's one of few period masters who does this, as far as I know. Fabris does a bit as well, but not nearly as exhaustively. Meyer, Capoferro, and other period masters sort of cover this, but not at the level of exhaustiveness that would be useful.
• Drills tend not to be designed cover this. The Capoferro Hierarchy Drill covers this somewhat, which is why it is such a good drill.
• This is the first layer that you can lie at. When a more experienced fencer does a half-lunge at a range that they can't stab a newer fencer at, they're lying here. They are telling the newer fencer that they will get stabbed, and the newer fencer, hapless as they are, believes it and jumps, giving the more experienced fencer the opportunity to stab them.
• Perception of these lies is what I believe to be the hallmark of a no-longer-beginner fencer. A lack of reaction to these lies is essential to fencing correctly, and is one of the most pernicious mistakes that even skilled fencers make. This isn't to say you shouldn't move at all - something that is a lie can also be a way to reposition for a different technique. And that can lead into jockeying for position. However, a twitch "HEY I'M THROWING A LUNGE" from out of measure is different, and responding to that is evidence of a deficiency in someone's fencing.
• I tend to call things that cover this "exercises" rather than drills. I think this is what most people use slow-fencing for. I tend to want to do them at-speed, because it's easy to accidentally react faster than possible when you're operating at 1/4 speed. Though, a case could be made for slow-fencing in that if your opponent predicts what you're doing and when you will do it, they could move that fast.
• A set of tactics, or the techniques you plan to engage in and the operations they can flow into.
• This is what I view as the highest useful level - a game-plan of what can be done. A good game-plan assumes that your opponent will do the thing that is the worst-for-you possible smart move that is based on them reacting to you or you reacting to them. Here, we are not yet thinking about predicting our opponent's action.
• This is what I forgot to have in mind yesterday.
• A basic Italian set of tactics would be something like this:
• Get in a backwards-leaning guard just out of both you and your opponent's lunge measure.
• If they manage to step forward and lunge or pass at this stage, execute a single-tempo or duo-tempi parry/riposte.
• Take a small step forward with your front foot and find their sword to the inside or outside, whichever occupies more space
• If they disengage, find their blade to the other side. Your hilt should be low enough that they can't strike to your body under your sword.
• Complete the small step with your back foot, leaning forward and progressing your find to a gain.
• If they disengage here, you can probably just lunge and stab them.
• Lunge and stab them through the eye.
• If they do an oh-shit emergency parry at this stage, they probably have to come off-line enough that their sword isn't a threat any more. Execute a tiny disengage around their hilt and stab them in the chest.
• That set of tactics is very basic, and doesn't address everything! It says nothing about if your opponent uses their off-hand or an off-hand implement, nor does it say things about if your opponent does weird things to gain your blade from below. It's a basic framework, and as time passes more things get hung from it. Maybe it has a deep strategic deficit which means the person using it will always lose if someone does a very particular sequence of actions. Who knows! Diagnosing these problems and searching through them is what I fucking live for, in this sport.
• Being a cold and ruthless killer means living at level and trying not to go to a higher level of tactics. This is the level at which things work well and consistently. Above here, we get to strange games of anticipation and then knowing your opponent. If you allow yourself to get sucked into those, it allows you to be lied to. If your opponent can firmly convince you that something is going to happen, then they do something else, then you will lose to them. It's a game of "who can lie better", and I think it's best not to play because there will always be a better liar.
• This is a very Spanish sentiment. The Spanish abhor lies and feints. I happen to agree with them philosophically, though I don't necessarily agree with their system.
• On the other side of things, Giganti and Capoferro wax rhapsodic about how the pinnacle of fencing is deceit. Many people agree with them, and do quite well with it. I will not say they're wrong, but it feels like a shallow end to the game. I'd rather work on perfecting my strategy, since that is universal.
• Fabris is the only person who really covers this in any depth, as far as I know. This is essentially all of what Fabris's Book Two is. He lists six game-plans with single rapier, and then four game-plans with sword and dagger. He then flow-charts out what you should do based on your opponent's reactions. I wish he had explicitly stated where these tactics don't work, and when to abandon them for other things. He says it in the positive sense, but I wish he said it in the negative sense more often. I understand that hubris is period, but still. 
•  Tactical Deceit
• This is the level at which you create and break expectations in people. It's very useful, and relies on finding quick rock-paper-scissors exchanges. Generally, this is implemented on the offense - I execute attack A, letting you execute defense A. Then I do it again. The third time I start off looking like I'm doing attack A, but then switch to attack B, which defeats defense A.
• Some basic patterns here include:
• A-A-B
• A-B-D
• In general, "do one thing until you don't do it"
• In general, "do a progression then skip a step in that progression"
• This can be done defensively, I guess? But it's much weirder, and relies on your opponent being more on-the-ball and taking your bait. This is something that Maija Soderholm talks about in her book, "The Liar, the Cheat, and the Thief", but I have not worked with that book enough to comment on whether the thing she speaks of is different from what I try to do.
• The way I try to do this is the German way - making the final technique a "masterstroke" that counterattacks the "expected" technique, but also defends against all other direct, single-tempo attacks. It's not the easiest to set up, but it allows us to implement this level while still staying true to the principles of the previous level. Even doing that though, this level should still be subsidiary to Tactics.
• You work on this by fighting a bunch of pickups, over and over again, forever. Soderholm's book has many drills to work on it, but I have not even attempted these drills.
• Personal Knowledge
• This is the level of "Oh, Remy is really good at in-fighting, so I should do this particular thing." Or "Lupold likes to snipe, so it'd be good if I bum-rush him." It's all about knowing who you are fighting.
• This is a level I do not like to rely on. It relies on using what you know of people to stab them. However, they might know that you know these things, so they might be expecting you to do a thing. But if you know that they know that you know, then you can do a different thing. But if they know that you know thaewnlkweanglakewnflkseanf newlafEWANFIAWNFLAENDFLKFNADFKLNFKLANDSF
• As stated, I don't like it because it can spiral infinitely. Now, if you can execute tactically sound operations which don't leave openings and Just Happen to strike at places you know your opponent is weak, you can do that. In fact, I encourage it. It should be subsidiary to Tactics, however.
• This all boils down to understanding what parts of The Game Of Fencing you understand better than your opponent. If you know this ahead of time, it might give you an edge. Might. In reality, I think it's best to just fight your fight and diagnose things from the flow of the fight, rather than from outside knowledge.
• Of course, if you are attempting to create winning Pennsic Champions pairings, this skill is super important. This skill is what lets you look at a fight and figure out who is more or less likely to win.
• Personal knowledge of yourself is important, though - it's good to know what parts of the game of fencing you are more or less strong at. In a tournament, it lets you attempt to lead bouts away from those areas. For example, if I'm not strong at in-fighting, I can stay at a distance. In training, it lets you decide what to beat your head against until you understand it better. This means improving every level below this one in that area of the game.
• In training, if you rely on this too much, then you and your opponent might, over the years, descend into a shallow sub-game of the overall game of fencing, in which you both attempt to hammer at a particular part of the game of fencing. This is why R&D is important - R&D is what happens when you try to break out of your known "best" fight and branch out, to try to find other techniques that are effective or useful in your overall game.

 So, that's the full breakdown of "What is Tactics", from my perspective. Hopefully it's useful to someone other than me. Hell, hopefully breaking it down like this is useful to me. The short version of the list is:

• Movements are smaller than,
• Techniques are smaller than,
• Operations are smaller than,
• Exchanges are smaller than,
• Tactics, which should be your focus above all,
• But if you can use Deceit you might as well,
• And if you can use Personal Knowledge then why not.

 I feel like I could do better. Like, making that into a poem? Seems like effort.

Bored of writing now, so I'm done.